Although a number of behavioral and pharmacological treatments for alcoholism have proven efficacy, there are several major challenges to the implementation and effectiveness of our current treatment models in community settings, including lack of interest in traditional specialty care models, high rates of early dropout in individuals who do enter traditional community programs, and considerable variability of response in patients who do become engage in treatment. Recent work done by our group and others suggests that adaptive treatment models can provide a solution to many of the problems that interfere with the successful delivery of effective alcoholism treatment. These models, which have also been referred to as "stepped care" and "treatment algorithms," are designed to increase participation in treatment by providing flexible care that is tailored over time on the basis of patient response to treatment. Adaptive approaches also seek to maximize participation by specifying lower intensity treatment whenever possible, and incorporating factors such as patient preference. The goal of this revised P20 is the further development of patient centered adaptive models of care for individuals with alcohol use disorders. The Research Methods Core will address a number of methodological issues in adaptive treatment, including identification of optimal tailoring variables that trigger modifications in treatment, statistical models for mediation, development of computer systems to implement complex adaptive algorithms, and identification of within-session therapeutic processes that facilitate adherence. The Principal Research Core consists of Developmental Projects that address several components of our conceptual model, including the development of adaptive models of care for patients identified with alcohol dependence in primary care settings;adaptive treatment algorithms for patients who enter community-based specialty care;and an innovative, web-based adaptive continuing care protocol. The impact of patient choice in treatment selection will be examined in two projects. The multidisciplinary P20 will facilitate collaborations between scientists with expertise in behavioral and pharmacological interventions, adaptive designs, statistics, economics, and dissemination, and directors of community-based programs, counselors and patients. Overall, the work of the P20 is expected to improve the feasibility, acceptability, and effectiveness of adaptive treatment designs.